1
Breeding for behavioural change in farm animals: Practical, economic and ethical
2
considerations
3
RB D’Eath1, J Conington1, AB Lawrence1, IAS Olsson2,3, P Sandøe3
4
5
1
6
2
7
8
3
Sustainable Livestock Systems, SAC, UK
IBMC- Instituto de Biologia Molecular e Celular, Universidade do Porto, Portugal
Danish Centre for Bioethics and Risk Assessment, University of Copenhagen, Denmark
9
10
[email protected]
11
12
Abstract
13
14
In farm animal breeding, behavioural traits are rarely included in selection programmes
15
despite their potential to improve animal production and welfare. Breeding goals have
16
been broadened beyond production traits in most farm animal species to include health
17
and functional traits, and opportunities exists to increase the inclusion of behaviour in
18
breeding indices.
19
20
On the technical level, breeding for behaviour presents some particular challenges
21
compared to physical traits. It is much more difficult and time-consuming to directly
22
measure behaviour in a consistent and reliable manner in order to evaluate the large
23
numbers of animals necessary for a breeding programme. For this reason, the
24
development and validation of proxy measures of key behavioural traits is often required.
25
Despite these difficulties, behavioural traits have been introduced by some breeders. For
26
example, ease of handling is now included in some beef cattle breeding programmes.
27
28
While breeding for behaviour is potentially beneficial, ethical concerns have been raised.
29
Since animals are adapted to the environment rather than the other way around, there may
30
be a loss of ‘naturalness’ and/or animal integrity. Some examples such as breeding for
1
31
good maternal behaviour could enhance welfare, production and naturalness, although
32
dilemmas emerge where improved welfare could result from breeding away from natural
33
behaviour. Selection against certain behaviours may carry a risk of creating animals
34
which are generally un-reactive (“zombies”), although such broad effects could be
35
measured and controlled. Finally breeding against behavioural measures of welfare could
36
inadvertently result in resilient animals (“stoics”) that do not show behavioural signs of
37
low welfare yet may still be suffering. To prevent this, other measures of the underlying
38
problem should be used, although cases where this is not possible remain troubling.
39
Keywords
40
Animal Welfare, Economics, Ethics, Genetics, Proxy measures, Selection index
41
Introduction
42
43
Breeding to change behaviour in farm animals has a number of possible benefits
44
including improving production and product quality, reducing labour costs and improving
45
handler safety (Jones & Hocking 1999; Boissy et al 2005; Grandinson 2005; Turner &
46
Lawrence 2007; Macfarlane et al 2009). Breeding for behaviour could also be used to
47
improve animal welfare, since many welfare problems may result from a mismatch
48
between the environment and animal’s range of coping responses (Fraser et al 1997).
49
Normally, animal welfare scientists try to identify ways to correct this mismatch by
50
changing the environment, although changing the animal by some means such as through
51
genetic selection (Muir & Craig 1998; Jones & Hocking 1999; Kanis et al 2004) is a
52
logical alternative.
53
54
Animal behaviour has undergone alteration throughout the history of domestication, and
55
at first this was not deliberate: only relatively docile members of a species could be
56
captured and/or herded, unmanageable animals were eaten rather than kept for breeding
57
(Price 1984; Mignon-Grasteau et al 2005). Over the centuries selection became more
58
deliberate, and is now carried out according to scientific principles in most farm animals,
59
primarily to ‘improve’ production traits. Initially, relatively few traits such as growth rate,
2
60
egg or milk yield were selected, but breeding goals have been refined by the addition of
61
further traits relating to efficiency (feed conversion efficiency), or product quality (lean
62
meat %, carcass composition, protein content of milk). In recent years, ‘functional’ traits
63
relating to health, biological functioning and longevity have come to be included
64
alongside traditional production traits in breeding indices, typically with an economic
65
weighting (Lawrence et al 2004).
66
67
In general there is growing interest in how breeding may affect animal welfare in a
68
negative or positive way. The Standing committee of the European convention for the
69
protection of animals kept for farming purposes which covers all major farmed species
70
(e.g. T-AP 1995; T-AP 1999; T-AP 2005a; T-AP 2005b) includes in its recommendations
71
an article on ‘changes of genotype’ which emphasises that breeding goals should include
72
health and welfare. Behavioural traits typically have heritability of a similar magnitude to
73
traits already included in breeding programmes, making it technically possible to include
74
behaviour, which is indeed already happening in some breeding programmes.
75
76
In this paper we will discuss a number of potential practical, economic and ethical issues
77
which affect the feasibility and desirability of genetic selection for behaviour. We begin
78
by outlining the process of animal breeding, introducing and defining concepts such as
79
heritability, genetic correlation and selection indices. We then introduce the evidence that
80
behaviour can be changed by genetic selection, discuss which behavioural traits have
81
been investigated at the genetic level in farm animals, and which of these have been
82
implemented in practice. We then describe some practical and economic factors affecting
83
implementation, and finally discuss some ethical considerations.
84
Modern livestock breeding
85
The scientifically-based breeding (quantitative genetics) used in most farm animal
86
species combines several desirable characteristics into a ‘breeding index’ or ‘selection
87
index’ of overall merit (Hazel 1943). The relative emphasis placed on each trait depends
88
on the other traits in the breeding objective. The rate of genetic change in a trait is
89
therefore determined by its heritability (defined below), its genetic correlation with other
3
90
traits in the index (defined below), the amount of variation seen in the population under
91
selection and the relative importance placed on the trait by the breeder (usually
92
determined by an overall breeding goal which is economic in the first instance).
93
94
The heritability of a trait can be described as the proportion of total variation that is
95
genetic (rather than environmental) in origin on a scale of 0 to 1, and is used to determine
96
an upper limit for how much genetic progress can be expected during selection. Traits
97
with a high heritability are usually more readily altered through selection. The genetic
98
correlation between two traits is a measure of the extent to which the same genes are
99
responsible for influencing both traits, on a scale of -1 to +1. Although it is easier to
100
make genetic progress with positively correlated traits, using selection index
101
methodology, it is possible to make progress with traits that are antagonistically
102
(unfavourably) correlated as long as the correlation between them is not close to 1.
103
Research into the genetics of behaviour
104
Behaviour is much more affected than physical traits by environmental influences either
105
at the time (e.g. presence of group-mates or humans) or in advance of behaviour (e.g.
106
learning or developmental influences). Nevertheless, there is still considerable evidence
107
for genetic influences on behaviour. This evidence comes from the existence of species
108
and breed differences, and studies involving quantitative genetics, artificial selection and
109
gene knock-out studies (reviewed by Reif & Lesch 2003; Mormède 2005; Van Oers et al
110
2005). The variety and extent of behavioural change that has been documented in
111
laboratory animal genetic studies (e.g. Miczek et al 2001; Finn et al 2003) indicates the
112
potential for similar genetic changes in behaviour in farm animals.
113
114
In farm animals, heritability has been estimated for a number of behavioural traits that are
115
of interest (most affect some aspect of production or welfare; Table 1). In many cases,
116
estimated heritabilities are of comparable magnitude to traits already included in breeding
117
programmes (around 0.1 to 0.4 REF), suggesting that selection for behaviour would be
118
possible in principle.
119
4
120
TABLE 1 ABOUT HERE
121
122
In addition to the individual behaviours outlined in Table 1, other authors have proposed
123
breeding goals which would be expected to affect more general aspects of behaviour.
124
Such approaches include breeding to reduce fearfulness (Jones & Hocking 1999; Boissy
125
et al 2005), stress reactivity (Mormède 2005), adaptability (Mignon-Grasteau et al 2005)
126
or robustness (Kanis et al 2004). Concerns have been raised about the risks of breeding
127
for traits with such wide effects (Mignon-Grasteau et al 2005).
128
129
A ‘group selection’ approach has been proposed as an indirect means to reduce negative
130
social behaviour between animals. The idea here is that conventional quantitative genetic
131
approaches can be altered to include the effect that animals have on each others’
132
production (Bijma et al 2007a; 2007b; Rodenburg et al 2009). In this way, negative
133
behaviours such as damaging behaviour (feather pecking, cannibalism, tail biting) or
134
aggressive behaviour (causing stress and excluding others from feeding) which affect
135
production variables (survival, growth, egg production) can be indirectly reduced.
136
137
For example, groups of laying hens were left with their beaks not trimmed and entire
138
groups were selected on the basis of longevity and egg production, resulting in lines
139
which did not require beak trimming (Muir & Craig 1998). Considerable mortality was
140
involved in this method which therefore should give rise to ethical concerns. A similar
141
methodology has been applied to pigs (Bergsma et al 2008; Canario et al 2008). The
142
actual effect on behaviour of applying this methodology can be assumed but as yet has
143
not been studied in much detail. It may be expected that the methodology will result in
144
general changes affecting more than one behaviour (Canario et al 2008; Rodenburg et al
145
2009).
146
147
5
148
Genetic selection for farm animal behaviour
149
For mink genetic research into various aspects of behaviour (exploration, fear of humans,
150
aggression, activity, stereotypy, pelt- and tail-biting; reviewed by Vinke et al 2002) has
151
shown that selection for behaviour is feasible; and selection experiments producing low
152
fear (Malmkvist & Hansen 2001) and low stereotypy (Svendsen et al 2007) have taken
153
place in Denmark. In Danish mink production animals are now selected against fur
154
chewing (Malmkvist & Hansen 2001) and in the Dutch production they are selected
155
against stereotypy and tail biting (Vinke et al 2002). The Standing committee of the
156
European convention for the protection of animals kept for farming purposes (T-AP
157
1999), now recommends that for fur animals: “Strongly fearful animals should not be
158
included in the breeding stock.”
159
160
Cattle may be dangerous to handle, and temperament in response to human handlers
161
(docility) has been used a criterion for genetic selection by the Limousin breed societies
162
in Ireland and Australia and is now being introduced in Britain (Irish Limousin Cattle
163
Society 2009; Australian Limousin Breeders Society 2009; British Limousin Cattle
164
Society 2009). The methods used vary, but in Ireland, a 1-10 scale (aggressive to docile)
165
is used depending on the response to a standard behavioural test in which a handler
166
attempts to move an animal to one corner of a pen and hold it there (Le Neindre et al
167
1995). In many countries, temperament is scored in dairy cattle and recorded for
168
inclusion in breeding indices. In the UK farmers rate their impressions of a cow on a 1-9
169
scale based on responses to milking (nervous to quiet, Pryce et al 2000).
170
171
Maternal behaviour in sheep (measured by a scoring system based around the proximity
172
to the lamb during tagging) has been shown to have a heritability of around 0.13 (Lambe
173
et al 2001). Efforts to improve this and other aspects of lamb vigour and maternal
174
behaviour around parturition are now being implemented in the UK sheep industry
175
(Conington et al 2009; Macfarlane et al 2009).
176
177
Although not actually selecting for behaviour, the change in breeding goal from litter size
178
at birth to litter size at day 5 in the Danish pig industry (Su et al 2007) is likely to have a
6
179
positive effect on aspects of maternal and neonatal behaviour that contribute to piglet
180
survival.
181
Practical issues affecting the implementation of selection for
182
behaviour
183
184
Measuring behaviour on the thousands of animals necessary to implement a breeding
185
programme raises a number of practical issues. The labour costs of measuring behaviour
186
by observation are high even for R&D, but are often prohibitive for practical
187
implementation. To reduce these costs, quick behavioural tests (e.g. 'stick' test in mink,
188
Malmkvist & Hansen 2001), automated measurement (e.g. flight speed from a crush in
189
Beef Cattle, Burrow 1997) or proxy traits (e.g. skin lesion number as a proxy for
190
aggressive behaviour; Turner et al 2006a; 2009a; 2009b) could be used. The use of proxy
191
traits as indicators of a more difficult-to-measure breeding goal trait is common practice
192
in breeding programmes (e.g. white blood cell counts in milk as an indirect indicator of
193
mastitis in dairy cows; Pryce et al 1998). Behavioural problems which occur in sudden
194
unpredictable outbreaks (e.g. hysteria, cannibalism and feather pecking in poultry; tail,
195
ear and flank biting in pigs) are particularly problematic to study. There is a need for
196
validated proxy measures that can be applied to animals in a ‘baseline’ state which are
197
predictive of their behaviour during an outbreak (e.g. Breuer et al 2001; Statham et al
198
2006).
199
200
There may however be unintended consequences of using proxy measures. For example,
201
breeding for slow flight speed in cattle in the hope of selecting calm animals might result
202
in animals which were slow for another reason (e.g. because they were lame), or that
203
breeding for few skin lesions 24hrs after mixing to reduce aggression in pigs could result
204
in blunt teeth rather than less fighting. To avoid these sorts of problems, the goal trait
205
must be clearly defined, and the genetic correlation between the goal and proxy trait
206
should be re-examined as breeding progresses.
207
7
208
Regardless of the recording method (behavioural observations, tests, scoring systems)
209
inter-observer reliability could be more of an issue for behaviour in comparison to simple
210
to measure traits such as weight or milk yield. This is especially a problem for multiple
211
farm breeding programmes where there is a single (different) scorer on each farm with
212
limited cross-checking (e.g. beef or sheep). In practice, even with these problems,
213
behaviour traits are heritable albeit at a low level (e.g. Pryce et al 2000). Poorly designed
214
scoring systems for behaviour, are likely to result in unexplained non-genetic sources of
215
variability in a trait and hence low heritability making it unlikely that a trait will be
216
adopted by breeders. Well designed, research-based objective scoring systems
217
(Macfarlane et al 2009) or (validated) use of automation (e.g. image analysis for feather
218
scoring or skin lesion scoring) provide potential solutions.
219
220
Potentially, the use of molecular markers or genome-wide selection could provide a cost-
221
effective way of selecting for behaviour, once the initial (expensive) research to identify
222
the genetic signature of a behaviour has been done (Désautés et al 2002; Mormède 2005;
223
Quilter et al 2007; Gutierrez-Gil et al 2008). However, as with any proxy trait, there is an
224
ongoing need to check the results against the actual behavioural phenotype for certain
225
animals every 2-3 generations. The genes or genome regions affecting differences in
226
behaviour are likely to vary with breed/country so there is a need for validation against
227
phenotype in each case.
228
229
Regardless of the trait and the method of measurement, genetic progress will be more
230
rapid if we better estimate the genetic component of variance; this is perhaps an
231
especially important point for behavioural selection given the sensitivity of behaviour to
232
short and long-term environmental influences. This requires environmental conditions to
233
be standardised or at least recorded (Mormède 2005) so that they can be included in the
234
statistical models used for genetic analysis.
235
Economic drivers and bottlenecks affecting the implementation of
236
selection for behaviour
237
8
238
In most farmed species, breeding goals are primarily aimed at production traits and the
239
relative weighting of traits in the selection index depends on their economic importance
240
(Brascamp et al 1985; Dekkers & Gibson 1998). There are a number of examples where
241
this has resulted in reduced welfare through unfavourable outcomes in health, welfare and
242
fitness characteristics, (see reviews by Rauw et al 1998; Jones & Hocking 1999; Sandøe
243
et al 1999). These traits were not recorded so the effects of breeding on them were
244
unknown or ignored. To address these problems, breeding goals have been broadened in a
245
number of species (e.g. sheep and dairy cows) to include more traits (Simm 1998;
246
Lawrence et al 2004; Pryce et al 2004).
247
248
It is important to note that many behavioural traits have an economic value. Thus by
249
analogy one reason to include health traits in Scandinavian dairy breeding is that for the
250
farmer, costs associated with mastitis (veterinary treatment, rejected milk) may offset the
251
gains from increased production (Christensen 1998). Although inclusion of behavioural
252
traits in breeding indices may constitute an improvement on animal welfare relative to not
253
including them, their inclusion at economically determined weights may only result in
254
slowing or halting in the growth of a problem, in particular if heritability is low or there is
255
unfavourable genetic correlation with other traits in the index (Nielsen et al 2006; Nielsen
256
& Amer 2007).
257
258
Some behavioural traits such as neonate survival or maternal behaviour may be of
259
sufficient economic weight to result in positive changes in animal welfare if
260
implemented. For other behavioural traits though, the economic value might be more
261
difficult to quantify, even though the outcomes might be desirable for farmers. For
262
example, large animals which are calm rather than reactive during handling could have
263
benefits for reduced labour costs, increased handler safety and meat quality (Turner &
264
Lawrence 2007) which are difficult to quantify in economic terms.
265
266
Society might wish behavioural traits to be improved more rapidly or even desire the
267
inclusion of some traits that enhance welfare at the expense of production (Olesen et al
268
2000; McInerney 2004). How could this be achieved? Methods to quantify the societal
9
269
benefits of broader breeding programmes and to estimate the non-market value of various
270
traits have been proposed (Olesen et al 2000; Nielsen et al 2006; Nielsen & Amer 2007).
271
Nevertheless, some traits will not have any economic value for the individual farmer, and
272
including them in the breeding goal may even come at an economic cost, as this slows
273
down the progress for traits that directly affect producer-income. Implementation of
274
breeding for such traits will only take place if special incentives are provided. Analogous
275
problems arise for other kinds of traits related to public goods such as reduced
276
environmental impact (Olesen et al 2000; Kanis et al 2005).
277
278
Rules to ensure animal welfare relating to animal transport, housing and slaughter
279
conditions are set by legislators, assurance schemes and retailers. Currently despite the
280
existence of recommendations on breeding by a number of bodies including FAWC
281
(2004), AEBC (2002) and the EU’s T-AP committee (e.g. T-AP 1995; T-AP 1999; T-AP
282
2005a; T-AP 2005b) there is, however, very little regulation of breeding goals (Lawrence
283
et al 2004). Existing EU legislation in this area has so far been ineffective (Olsson et al
284
2006).
285
286
Decision making over breeding goals varies according to the species involved. In pigs
287
and poultry, a few global breeding companies control breeding and determine the
288
breeding goals (in response to customer needs). Dairy cattle breeding is much more
289
diverse in terms of ownership of pedigree animals, although genetic evaluations are
290
centralised. Estimated breeding values for each bull for each trait are published, allowing
291
farmers (to some extent) to make decisions about which traits to focus on when
292
purchasing semen.
293
294
In the UK sheep and beef industries, some farmers make use of schemes which enable
295
breeding index methodology to be applied to systematically improve certain traits, but a
296
substantial number of pedigree breeders do not. Thus, there is for these breeds some room
297
not only for breeding organizations but also for individual farmers to consider additional
298
traits other than production traits in breeding.
299
10
300
In the EU, there has been an initiative of self-regulation by breeders: the Code of Good
301
Practice for European Farm Animal Breeding and Reproduction (CODE-EFABAR,
302
Neeteson-van Nieuwenhoven et al 2006) and some voluntary engagement by individual
303
breeding companies with ethicists (Olsson et al 2006).
304
305
Presently under schemes such as organic, Freedom Foods or Products of Protected
306
Origin, consumers pay premium prices for products with perceived added value in terms
307
of production system. However, as opposed to production systems, consumers are
308
unlikely to be aware of the role of breeding, and it being such a small part of the
309
production process will probably make it difficult to justify a price increase (Olsson et al
310
2006). This may however be different if existing labelling schemes would also
311
incorporate breeding as part of their requirements. At present, this is only done indirectly,
312
as when assurance programs require animals of a certain breed such as slow-growing
313
broilers (Cooper & Wrathall 2009) or locally adapted animals.
314
315
Ethical issues arising from selection for behaviour
316
317
Many people feel that limits should be placed on our interference with nature (Banner
318
1995; AEBC 2002; Macnaghten 2004). And it should be expected that this feeling might
319
be strong in cases where we are tangling with complex aspects of animals’ natures such
320
as the genetic basis for their behaviour. Along with an animal’s feelings and state of
321
health (Fraser et al 1997), the opportunity to express normal (or natural) behaviour is seen
322
as an important aspect of animal welfare, and it is one of FAWC’s five freedoms (FAWC
323
2004).
324
325
The call for ethical limits can be defended in two rather different ways. It can be claimed
326
either that we should refrain from interfering because we cannot accurately foresee the
327
consequences of what we are doing and may therefore bring about some kind of disaster,
328
or alternatively that we should leave nature as it is because untouched nature has a value
329
of its own (Banner 1995; AEBC 2002; Macnaghten 2004).
11
330
331
According to the first line of thought the problem with interfering is that we cannot
332
properly predict the long-term consequences of what we are doing. If we try to
333
manipulate nature on the basis of ‘grand plans’ for the future, there is a real danger that
334
unexpected and harmful consequences occur – as indeed it has sometimes happened for
335
example when species of animals have been introduced by humans in new territory.
336
337
According to the other line of thought the problem with interfering with nature is that we
338
should respect what is seen as the integrity of nature. It is seen as perverse and wrong that
339
we try to shape animals according to our plans rather than leaving them to be the kind of
340
creatures they are. Of course, in the context of farm animal breeding it may sound a bit
341
weird to appeal to the idea that it is wrong to change animals to fit our goals – since that
342
in a way is the raison d’etre of animal breeding. However, some argue that integrity
343
comes in degrees and that it is a bigger concern to manipulate the behaviour of a dairy
344
cow than it is to manipulate its disease resistance or length of calving intervals (Siipi
345
2008).
346
347
Changing the holes or the pegs?
348
349
Animal welfare problems often result where there is a mismatch between an animal’s
350
coping ability and the range of challenges offered by the environment (Fraser et al 1997).
351
Bernard Rollin (2002) has characterised intensively farmed animals as square pegs forced
352
into round holes; and breeding to make the animals fit the environment may be seen as an
353
attempt to change the pegs rather than the holes.
354
355
Changing the environment to suit the animal is usually seen as the solution, but why is
356
this ethically preferable to changing the animal to suit the environment? Concerns over
357
animal naturalness or integrity are the issue here. In addition, since biology appears to
358
impose few limitations on what is possible, changing the animal to suit the environment
359
raises the question of the ethical acceptability of the environment. In a discussion of how
360
breeding could be used to improve pig welfare, Kanis et al (2004) recognised that
12
361
breeding animals adapted to tolerate poor environments might result in a decline in
362
housing or husbandry practices.
363
364
To address this problem, Lawrence et al (2001) proposed that we should begin by
365
defining ‘Ethical Environment Envelopes’ and then breed animals to have good welfare
366
within these. There are a number of examples where breeding for behaviour could suit
367
animals to more extensive housing systems which may be viewed as ethically more
368
desirable than the alternative intensive housing systems. For example, selection for good
369
maternal and neonatal behaviour in pigs could facilitate a move away from confinement
370
housing, and selection to reduce feather pecking in barn and free-range laying hens will
371
make the move away from cages easier and might reduce the need for beak trimming.
372
Similarly, extensive systems for sheep could be made easier by breeding for animals that
373
are disease resistant and do not require shearing, tail docking or close supervision at
374
lambing (Conington et al 2009).
375
376
In intensive systems, even though it is more controversial, an argument could be made for
377
pragmatism and accepting genetic selection for behaviour as part of the solution for
378
welfare problems. For example, tail-biting in pigs could be reduced by the provision of
379
more space and particularly improved access to substrates for rooting and chewing.
380
However, the vast majority of pig farms in the EU do not provide adequate substrates and
381
painful tail docking is widely applied. Tail docking removes the welfare problem for the
382
bitten pig, but not for the biter- it simply masks the fact that these pigs still lack a suitable
383
outlet for their motivation to root and chew on something.
384
385
Selection to reduce tail biting is ethically less attractive than providing suitable substrates,
386
since it compromises the pig’s integrity, particularly if accompanied by a correlated
387
reduction in other behaviours which could be seen as being central to ‘pigness’ such as
388
rooting and chewing. On the other hand if the alternative is tail docking breeding to
389
reduce tail biting may be seen as the smaller of two evils. Thus a balance needs to be
390
struck. If we accept that pigs are going to continue to be kept in systems without suitable
13
391
substrates, then should we select against tail biting to improve pig welfare and removing
392
the need for tail docking at the risk of compromising the pigs’ integrity?
393
394
To take a different example, are we content with the ‘unnatural wolf’ (the dog) which is
395
happier in a domestic setting because it has no desire to hunt? Isn’t this better than a
396
‘natural’ wolf-like dog which is prevented from hunting? Of course, it may be argued that
397
much effort is put into ensuring that dogs live reasonable lives; whereas breeding against
398
tail biting in pigs could be seen as a too easy solution to the problem.
399
400
Zombies
401
One specific scenario, of particular concern to those concerned with animal integrity, is
402
that animals may become extremely inactive or generally un-reactive to external stimuli
403
as a result of breeding for behaviour. To simplify matters let us call these animals
404
“zombies”.
405
406
Reduced responsiveness to humans in particular (docility) and to environmental stimuli in
407
general has been a major feature of behavioural change throughout domestication (Price
408
1984), so further change in this direction could be thought of as purely a continuation of
409
the domestication process (Jones & Hocking 1999). Some authors have proposed
410
selection for animals that are less reactive to stress or less fearful across a wide range of
411
situations (Jones & Hocking 1999; Mignon-Grasteau et al 2005), and the ‘zombie’
412
criticism would apply to this kind of breeding. Indeed, Mignon-Grasteau et al. (2005)
413
acknowledge the need for an ethical debate in wider society before such proposals could
414
be taken forward. Even when a single trait is the focus of selection, genetic correlations
415
between traits mean that the impact could be wider: Pigs which were genetically less
416
aggressive at mixing were also less reactive at weighing (D'Eath et al 2009; Turner et al
417
2009a).
418
419
The issue of Zombies is clearly a problem for those advocating animal integrity. But why
420
should it matter from the point of view of an animal that it has a smaller number of
421
preferences and desires – as long as the desires that the animal does have are being
14
422
satisfied? After all isn’t animal welfare all about making sure that there is a fit between
423
what an animal needs or prefers and what it gets? (Sandøe 1996)
424
425
To answer this question one may seek inspiration from the utilitarian philosopher John
426
Stuart Mill (1863) who argued that “It is better to be a human being dissatisfied than a
427
pig satisfied; better to be Socrates dissatisfied than a fool satisfied. And if the fool, or the
428
pig, are a different opinion, it is because they only know their own side of the question."
429
The idea would be that breeding zombie animals is problematic because it means
430
reducing the value of the animal lives that comes out of the process.
431
432
The thought experiment of deliberately breeding animals with reduced sentience (a
433
reduced capacity for higher mental states) was considered in the Banner report (Banner
434
1995) as being ‘objectionable in its own right’. Others have expressed concern that
435
reduced sentience could inadvertently result from selection for behavioural change
436
(Paragraph 110, FAWC 2004).
437
438
Of course, since animal sentience is difficult to prove or measure it is difficult to address
439
these issues in practice. However even in theory they may be a disagreement between
440
those who think that animal welfare is all about making sure that animals get what they
441
need and want and those who think that a higher level of needs and wants makes room for
442
a richer and better life. The authors of this paper tend to side with the former.
443
444
Stoics
445
A very different scenario from the one just discussed is that animals are being bred to
446
change behaviour, but they still experience the negative feelings associated with the
447
unwanted behaviour. These animals we shall here call stoics, because outward signs of
448
suffering appear to be reduced. This scenario could perhaps be thought of as falling
449
within the ethical concern of ‘unintended consequences’.
450
451
In relation to disease or parasitism, the concepts of resistance and resilience have subtly
452
different meanings. Resistant animals do not become infected at all, while resilient
15
453
animals are able to function better (growing and reproducing) despite being infected
454
(Albers et al 1987). If one were to infect a population and just measure growth rate, these
455
two classes of animals might appear similar, while from a welfare perspective resistance
456
is surely preferable to resilience.
457
458
An analogous situation could occur when breeding to change behaviour, where stoics
459
could be thought of as similar to resilient animals. Genetic selection directly on a trait
460
which is used to measure welfare might mean that the trait becomes a less reliable
461
indicator of welfare: A thought experiment here might be, that selection to improve
462
locomotion score in lame animals could result in animals which still have the underlying
463
problem (with bad feet or joint problems) but which do not show it. Selection to change
464
behaviour without understanding the mechanism of that change could result in the mental
465
equivalent of lameness (e.g. high fearfulness could result in inactivity).
466
467
Whenever possible, direct examination of the source of the problem is important to
468
prevent such undesired effect (e.g. Conington et al 2009). However, as illustrated by the
469
discussion around the example provided by Mills and co-workers (Mills et al 1985a;
470
1985b), this may not be straightforward. These researchers reduced stereotypic pacing
471
behaviour in poultry by selecting against the amount of pre-laying pacing. Mason et al
472
(2007) argued that this would be more likely to result in an improvement in welfare than
473
selection against the stereotypy itself, because pre-laying pacing was an indicator of
474
motivation to find a nest, so the root cause of the stereotypy had been altered. However,
475
Appleby and Hughes (1991) argued that it had not been established whether reduced pre-
476
laying pacing indicated that these animals actually experienced less frustration in the
477
absence of a nest.
478
479
Muir and Craig (1998) describe another example: “Duncan and Filshie (1980) showed
480
that a flighty strain of birds that exhibited avoidance and panic behaviour following
481
stimulation returned to a normal heart beat sooner than a line of more docile birds,
482
implying that the docile birds may be too frightened to move”. Different species of
483
penguins (in the wild) differ in their behavioural reactivity to approaching humans
16
484
(Holmes 2007), but even penguins who show little behavioural reaction may show
485
prolonged elevations in heart rate, suggesting that they experience an emotional response
486
(Nimon et al 1995; Ellenberg et al 2006). Thus the link between emotional state and
487
outward behaviour is not straightforward and must be understood before beginning on a
488
selection programme to change behaviour.
489
490
When a welfare end-point, such as the level of stereotypic behaviour is directly selected
491
against (e.g. by the mink industry in the Netherlands; Vinke et al 2002), this could
492
present an example of selecting only against the symptoms while masking an underlying
493
problem (Mason et al 2007). Indeed high stereotyping mink often have lower endocrine
494
stress responses than low stereotyping mink, suggesting that it is a successful coping
495
mechanism (Mason & Latham 2004). Svendsen et al (2007) found that low stereotypy
496
was associated with high levels of fear of humans.
497
498
Kanis et al (2004) propose that experiments in which animals learn a task to express their
499
environmental preferences could be used in selection. “It could be a practical option to
500
breed for pigs which are less motivated to improve or change their situation and are thus
501
sufficiently satisfied”. There is a risk that this approach might result in stoical pigs which
502
do not act to remove themselves from stress, apathetic inactive pigs, or even those which
503
are poor at learning such tasks.
504
505
There is thus some technical support for this ethical concern of ‘meddling with what we
506
don’t understand”. For example, Mormède (2005) in a review of the opportunities to use
507
molecular genetics in breeding for behaviour states that “However, a major limit to these
508
studies is the limited basic knowledge about psycho-biological dimensions underlying
509
behavioural trait variability, and the availability of reliable and meaningful measures of
510
these,”
511
512
In summary, we believe that this issue that we have discussed under the heading of
513
‘stoics’ represents a real ethical issue, where an illusion of improved welfare might mask
514
a continuing underlying problem, such as thwarted motivation.
17
515
516
Conclusions and Animal Welfare Implications
517
We have argued that breeding to change behaviour offers potential for improving
518
production and welfare. It is technically possible to do, although there are various
519
practical issues that need to be addressed for successful implementation. Primarily there
520
is a need for well-validated abbreviated methods of recording behaviour or its proxies.
521
Economics as the key driver for breeders will always be a barrier to implementation of
522
behavioural traits relating to non-economic welfare traits, although there are of course a
523
number of win-win traits where there is less conflict between profit and welfare such as
524
reducing neonatal mortality.
525
526
Ethical concerns over ‘meddling with nature’ when breeding for behaviour need to be
527
considered. In particular the issues of unforeseen consequences of selection (for example
528
due to antagonistic genetic correlations between traits) and the reduction of animal
529
integrity or naturalness are important. In terms of naturalness, domesticated animals are
530
already compromised in this regard, making clear-cut definitions difficult. Where the
531
environment for which selection occurs is seen as ethically desirable (e.g. extensive, free-
532
range), there may be fewer problems, but decisions over selection to change behaviour in
533
intensive environments could involve balancing between opportunities to improve animal
534
welfare and the risk of reduced animal integrity.
535
536
Our position is that the resulting animal welfare (animal feelings) is of paramount
537
importance here. The specific concern that selection for behaviour could result in
538
extremely docile “Zombies” may give rise to disagreement between those who like the
539
authors of the present paper are mainly concerned about preventing welfare problems for
540
the animals and those who care about animal integrity and see excessively docile animals
541
as lacking something of significant value. Breeding for “Zombies” could be guarded
542
against in a selection programme by ensuring that a variety of behaviours are recorded,
543
and the genetic correlations among them and other breeding goals are understood. A
544
more important concern is the issue of “stoical” animals where breeding against
18
545
behavioural (or other) indicators of welfare could mask a problem without really solving
546
it, unless great care is taken in identifying accurate measures of the underlying problem,
547
which may not always be possible when unobservable mental states are the ultimate
548
indicator of a problem.
549
550
Acknowledgments
551
SAC is supported by the Scottish Government.
552
553
References
554
555
AEBC 2002 Animals and Biotechnology, a report by the Agriculture and Environment
Biotechnology Commission.
556
557
558
Albers GAA, Gray GD, Piper LR, Barker JSF, Lejambre LF and Barger IA 1987
The Genetics of Resistance and Resilience to Haemonchus-Contortus Infection in Young
Merino Sheep. International Journal for Parasitology 17: 1355-1363
559
560
561
Appleby MC and Hughes BO 1991 Welfare of laying hens in cages and alternative
systems: environmental, physical and behavioural aspects. World's Poultry Science
Journal 47: 109-128
562
563
Australian Limousin Breeders Society 2009 Limousin Assurance Programme.
http://www.limousin.com.au/assurance-program.html .
564
565
566
Banner M 1995 Report of the Committee to Consider the Ethical Implications of
Emerging Technologies in the Breeding of Farm Animals. London: HMSO, Ministry of
Agriculture, Fisheries and Food.
567
568
Beilharz RG and Cox DF 1967 Genetic analysis of open field behavior in swine. Journal
of Animal Science 26: 988-990
569
570
571
Bergsma R, Kanis E, Knol EF and Bijma P 2008 The contribution of social effects to
heritable variation in finishing traits of domestic pigs (Sus scrofa). Genetics 178: 15591570
572
573
Bijma P, Muir WA and van Arendonk JAM 2007a Multilevel selection 1: Quantitative
genetics of inheritance and response to selection. Genetics 175: 277-288
574
575
576
Bijma P, Muir WM, Ellen ED, Wolf JB and van Arendonk JAM 2007b Multilevel
selection 2: Estimating the genetic parameters determining inheritance and response to
selection. Genetics 175: 289-299
19
577
578
Boissy A, Fisher AD, Bouix J, Hinch GN and Le Neindre P 2005 Genetics of fear in
ruminant livestock. Livestock Production Science 93: 23-32
579
580
Brascamp EW, Smith C and Guy DR 1985 Derivation of Economic Weights from
Profit Equations. Animal Production 40: 175-180
581
582
583
Breuer K, Sutcliffe MEM, Mercer JT, Rance KA, O'Connell NE, Sneddon IA and
Edwards SA 2005 Heritability of clinical tail-biting and its relation to performance traits.
Livestock Production Science 93: 87-94
584
585
586
587
Breuer KB, Beattie VE, Dunne LM, Slade EC, Davies Z, Mercer JT, Rance KA,
Sneddon IA, Sutcliffe MEM and Edwards SA 2001 Validation and development of a
behavioural test to predict the predisposition of growing pigs to perform harmful social
behaviour such as tail biting.pp. 50 BSAS Scarborough,
588
589
British Limousin Cattle Society 2009 Limousin Society to Introduce Docility Scoring.
http://www.limousin.co.uk/projects/docility_scoring.html .
590
591
Buitenhuis AJ and Kjaer JB 2008 Long term selection for reduced or increased pecking
behaviour in laying hens. Worlds Poultry Science Journal 64: 477-487
592
593
Burrow HM 1997 Measurements of temperament and their relationships with
performance traits in beef cattle. Animal Breeding Abstracts 65: 477-495
594
595
596
597
598
Canario L, Bergsma R, D'Eath RB, Lawrence AB, Roehe R, Lundeheim N,
Rydhmer L, Knol E and Turner SP 2008 Genetic relations between the group effect for
average daily gain, and post-mixing aggression and skin lesions in Swedish pigs.
200.Vilnius, Lithuania, Proceedings of the 59th Annual Meeting of the European
Association for Animal Production.
599
600
601
Christensen LG 1998 Possibilities for genetic improvement of disease resistance,
functional traits and animal welfare. Acta Agriculturae Scandinavica Section A- Animal
Science Suppl 29: 77-89
602
603
Conington J, Collins J and Dwyer CM 2009 Breeding for easier managed sheep.
Animal Welfare this issue:
604
605
Cooper MD and Wrathall JHM 2009 Assurance schemes as a tool to tackle genetic
welfare problems in broilers. Animal Welfare this issue:
606
607
Craig JV and Muir WM 1989 Fearful and Associated Responses of Caged White
Leghorn Hens - Genetic Parameter Estimates. Poultry Science 68: 1040-1046
608
609
Craig JV and Muir WM 1993 Selection for Reduction of Beak-Inflicted Injuries
Among Caged Hens. Poultry Science 72: 411-420
610
611
Craig JV, Ortman LL and Guhl AM 1965 Genetic selection for social dominance
ability in chickens. Animal Behaviour 13: 114-131
20
612
613
614
D'Eath RB, Roehe R, Turner SP, Ison SH, Farish M, Jack MC and Lawrence AB
2009 Genetics of animal temperament: aggressive behaviour at mixing is genetically
associated with the response to handling in pigs. Animal in press:
615
616
Dekkers JCM and Gibson JP 1998 Applying breeding objectives to dairy cattle
improvement. Journal of Dairy Science 81: 19-35
617
618
619
620
Désautés C, Bidanel JP, Milan D, Iannuccelli N, Amigues Y, Bourgeois F, Caritez
JC, Renard C, Chevalet C and Mormède P 2002 Genetic linkage mapping of
quantitative trait loci for behavioral and neuroendocrine stress response traits in pigs.
Journal of Animal Science 80: 2276-2285
621
622
623
Duncan IJH and Filshie JH 1980 The use of radio telemetry devices to measure
temperature and heart rate in domestic fowl. In: CJ Amlaner & DW Macdonald (eds) A
Handbook on Biotelemetry and Radio Tracking pp. 579-588 Pergamon Oxford,
624
625
626
Ellenberg U, Mattern T, Seddon PJ and Jorquera GL 2006 Physiological and
reproductive consequences of human disturbance in Humboldt penguins: The need for
species-specific visitor management. Biological Conservation 133: 95-106
627
628
629
Faure JM and Mills AD 1998 Improving the Adaptability of Animals by Selection. In:
T Grandin (ed) Genetics and the Behavior of Domestic Animals pp. 235-264 Academic
Press San Diego,
630
631
FAWC 2004 Report on the Welfare implications of Animal Breeding and Breeding
Technologies in Commercial Agriculture. London, Farm Animal Welfare Council. 2004.
632
633
Finn DA, Rutledge-Gorman MT and Crabbe JC 2003 Genetic animal models of
anxiety. Neurogenetics 4: 109-135
634
635
Fraser D, Weary DM, Pajor EA and Milligan BN 1997 A scientific conception of
animal welfare that reflects ethical concerns. Animal Welfare 6: 187-205
636
637
Grandinson K 2005 Genetic background of maternal behaviour and its relation to
offspring survival. Livestock Production Science 93: 43-50
638
639
Grandinson K, Rydhmer L, Strandberg E and Thodberg K 2003 Genetic analysis of
on-farm tests of maternal behaviour in sows. Livestock Production Science 83: 141-151
640
641
642
Gutierrez-Gil B, Ball N, Burton D, Haskell M, Williams JL and Wiener P 2008
Identification of Quantitative Trait Loci Affecting Cattle Temperament. Journal of
Heredity 99: 629-638
643
644
Hansen CPB 1993a Stereotypies in Ranch Mink - the Effect of Genes, Litter Size and
Neighbors. Behavioural Processes 29: 165-178
21
645
646
647
648
649
Hansen SW 1993b Selection for behavioural characteristics in mink and the effect on
reproduction results. Genetics, Breeding and Reproduction of Fur Animals, NJF Report
No 90, Proceedings of a NJFworkshop,
Viborg, Denmark 29 April pp. 35-41 Scandinavian Association of Agricultural Scientists
(NJF) Viborg, Denmark,
650
651
Hazel LN 1943 The genetic basis for constructing selection indexes. Genetics 28, 476490.
652
653
654
Hemsworth PH, Barnett JL, Treacy D and Madgwick P 1990 The Heritability of the
Trait Fear of Humans and the Association between This Trait and Subsequent
Reproductive- Performance of Gilts. Applied Animal Behaviour Science 25: 85-95
655
656
Holmes ND 2007 Comparing king, gentoo, and royal penguin responses to pedestrian
visitation. Journal of Wildlife Management 71: 2575-2582
657
658
Irish Limousin Cattle Society 2009 Irish Limousin Docility Reaching New Heights.
http://www.irishlimousin.com/docilityfile.html .
659
660
661
Jeppesen LL, Heller KE and Bildsoe A 2004 Stereotypies in female farm mink
(Mustela vison) may be genetically transmitted and associated with higher fertility due to
effects on body weight. Applied Animal Behaviour Science 86: 137-143
662
663
Jones RB and Hocking PM 1999 Genetic selection for poultry behaviour: Big bad wolf
or friend in need? Animal Welfare 8: 343-359
664
665
666
667
Kadel MJ, Johnston DJ, Burrow HM, Graser HU and Ferguson DM 2006 Genetics
of flight time and other measures of temperament and their value as selection criteria for
improving meat quality traits in tropically adapted breeds of beef cattle. Australian
Journal of Agricultural Research 57: 1029-1035
668
669
Kanis E, De Greef KH, Hiemstra A and van Arendonk JAM 2005 Breeding for
societally important traits in pigs. Journal of Animal Science 83: 948-957
670
671
672
Kanis E, van den Belt H, Groen AF, Schakel J and De Greef KH 2004 Breeding for
improved welfare in pigs: a conceptual framework and its use in practice. Animal Science
78: 315-329
673
674
Kjaer JB and Sorensen P 1997 Feather pecking behaviour in White Leghorns, a genetic
study. British Poultry Science 38: 333-341
675
676
677
Lambe NR, Conington J, Bishop SC, Waterhouse A and Simm G 2001 A genetic
analysis of maternal behaviour score in Scottish Blackface sheep. Animal Science 72:
415-425
678
679
Lawrence AB, Conington J and Simm G 2004 Breeding and animal welfare: practical
and theoretical advantages of multi-trait selection. Animal Welfare 13: S191-S196
22
680
681
682
Lawrence AB, Pryce JE and Simm G 2001 G x EEE: the missing link when breeding
for welfare. In: JP Garner, JA Mench & SP Heekin (eds)pp. 90 Center for Animal
Welfare UC Davis Davis, USA,
683
684
685
Le Neindre P, Trillat G, Sapa J, Menissier F, Bonnet JN and Chupin JM 1995
Individual-Differences in Docility in Limousin Cattle. Journal of Animal Science 73:
2249-2253
686
687
688
Løvendahl P, Damgaard LH, Nielsen BL, Thodberg K, Su GS and Rydhmer L 2005
Aggressive behaviour of sows at mixing and maternal behaviour are heritable and
genetically correlated traits. Livestock Production Science 93: 73-85
689
690
Macfarlane JM, Matheson SM and Dwyer CM 2009 Genetic parameters for birth
difficulty, lamb vigour and lamb sucking ability in suffolk sheep. Animal Welfare .
691
692
693
Macnaghten P 2004 Animals in their nature: A case study on public attitudes to animals,
genetic modification and 'nature'. Sociology-the Journal of the British Sociological
Association 38: 533-551
694
695
Malmkvist J and Hansen SW 2001 The welfare of farmed mink (Mustela vison) in
relation to behavioural selection: A review. Animal Welfare 10: 41-52
696
697
698
Mason G, Clubb R, Latham N and Vickery S 2007 Why and how should we use
environmental enrichment to tackle stereotypic behaviour? Applied Animal Behaviour
Science 102: 163-188
699
700
Mason G and Latham N 2004 Can't stop, won't stop: Is stereotypy a reliable animal
welfare indicator. Animal Welfare 13: S57-S69
701
702
McInerney J 2004 Animal welfare, economics and policy. Report on a study undertaken
for the Farm & Animal Health Economics Division of Defra.
703
704
Miczek KA, Maxson SC, Fish EW and Faccidomo S 2001 Aggressive behavioral
phenotypes in mice. Behavioural Brain Research 125: 167-181
705
706
707
Mignon-Grasteau S, Boissy A, Bouix J, Faure JM, Fisher AD, Hinch GN, Jensen P,
Le Neindre P, Mormede P, Prunet P, Vandeputte M and Beaumont C 2005 Genetics
of adaptation and domestication in livestock. Livestock Production Science 93: 3-14
708
Mill JS 1863 Utilitarianism. Longmans, Green, Reader, and Dyer: London
709
710
Mills AD, Duncan IJH, Slee GS and Clark JSB 1985a Heart-Rate and Laying
Behavior in 2 Strains of Domestic Chicken. Physiology & Behavior 35: 145-147
711
712
713
Mills AD and Faure JM 1991 Divergent Selection for Duration of Tonic Immobility and
Social Reinstatement Behavior in Japanese-Quail (Coturnix-Coturnix-Japonica) Chicks.
Journal of Comparative Psychology 105: 25-38
23
714
715
716
Mills AD, Woodgush DGM and Hughes BO 1985b Genetic-Analysis of Strain
Differences in Pre-Laying Behavior in Battery Cages. British Poultry Science 26: 187197
717
718
Mormède P 2005 Molecular genetics of behaviour: research strategies and perspectives
for animal production. Livestock Production Science 93: 15-21
719
720
Muir WM and Craig JV 1998 Improving animal well-being through genetic selection.
Poultry Science 77: 1781-1788
721
722
723
Neeteson-van Nieuwenhoven AM, Merks J, Bagnato A and Liinamo AE 2006
Sustainable transparent farm animal breeding and reproduction. Livestock Science 103:
282-291
724
725
Nielsen HM and Amer PR 2007 An approach to derive economic weights in breeding
objectives using partial profile choice experiments. Animal 1: 1254-1262
726
727
Nielsen HM, Christensen LG and Odegard J 2006 A method to define breeding goals
for sustainable dairy cattle production. Journal of Dairy Science 89: 3615-3625
728
729
730
Nielsen UL and Therkildsen N 1995 Selektionsforsøg for og imod pelsgnav hos mink.
In: N Therkildsen (ed) Faglig Årsberetning 1993/94 pp. 209 Pelsdyrerhvervets Forsøgs og Rådgivningsvirksomhed A/S: Holsterbro, Denmark,
731
732
Nimon AJ, Schroter RC and Stonehouse B 1995 Heart-Rate of Disturbed Penguins.
Nature 374: 415
733
734
Olesen I, Groen AF and Gjerde B 2000 Definition of animal breeding goals for
sustainable production systems. Journal of Animal Science 78: 570-582
735
736
737
Olsson IAS, Gamborg C and Sandoe P 2006 Taking ethics into account in farm animal
breeding: What can the breeding companies achieve? Journal of Agricultural &
Environmental Ethics 19: 37-46
738
739
740
Phocas F, Boivin X, Sapa J, Trillat G, Boissy A and Le Neindre P 2006 Genetic
correlations between temperament and breeding traits in Limousin heifers. Animal
Science 82: 805-811
741
742
Price EO 1984 Behavioral aspects of animal domestication. The Quarterly Review of
Biology 59: 1-32
743
744
745
Pryce JE, Coffey MP and Brotherstone S 2000 The genetic relationship between
calving interval, body condition score and linear type and management traits in registered
Holsteins. Journal of Dairy Science 83: 2664-2671
746
747
748
Pryce JE, Conington J, Sørensen P, Kelly HRC and Rydhmer L 2004 Breeding
Strategies for Organic Livestock. In: M Vaarst, S Roderick, V Lund & W Lockeretz (eds)
Animal Health and Welfare in Organic Agriculture pp. 357-388 CABI Wallingford, UK,
24
749
750
751
Pryce JE, Esslemont RJ, Thompson R, Veerkamp RF, Kossaibati MA and Simm G
1998 Estimation of genetic parameters using health, fertility and production data from a
management recording system for dairy cattle. Animal Science 66: 577-584
752
753
754
755
Quilter CR, Blott SC, Wilson AE, Bagga MR, Sargent CA, Oliver GL, Southwood
OI, Gilbert CL, Mileham A and Affara NA 2007 Porcine Maternal Infanticide as a
model for Puerperal Psychosis. American Journal of Medical Genetics Part BNeuropsychiatric Genetics 144B: 862-868
756
757
758
Rauw WM, Kanis E, Noordhuizen-Stassen EN and Grommers FJ 1998 Undesirable
side effects of selection for high production efficiency in farm animals: a review.
Livestock Production Science 56: 15-33
759
760
Reif A and Lesch KP 2003 Toward a molecular architecture of personality. Behavioural
Brain Research 139: 1-20
761
762
763
Rodenburg TB, Bijma P, Ellen ED, Bergsma R, de Vries S, Bolhuis JE, Kemp B and
van Arendonk JAM 2009 Breeding amiable animals? Improving farm animal welfare by
including social effects into the genetic model. Animal Welfare this issue:
764
765
766
Rodenburg TB, Buitenhuis AJ, Ask B, Uitdehaag KA, Koene P, van der Poel JJ and
Bovenhuis H 2003 Heritability of feather pecking and open-field response of laying hens
at two different ages. Poultry Science 82: 861-867
767
768
Rollin BE 2002 An ethicist's commentary on equating productivity and welfare.
Canadian Veterinary Journal 43[2], 83.
769
770
Sandøe P 1996 Animal and human welfare - are they the same kind of thing? Acta
Agriculturae Scandinavica Section A Suppl 27: 11-15
771
772
Sandøe P, Nielsen BL, Christensen LG and Sorensen P 1999 Staying good while
playing god- The ethics of breeding farm animals. Animal Welfare 8: 313-328
773
Siipi H 2008 Dimensions of naturalness. Ethics & the Environment 13: 71-103
774
775
Silva B, Gonzalo A and Canon J 2006 Genetic parameters of aggressiveness, ferocity
and mobility in the fighting bull breed. Animal Research 55: 65-70
776
Simm G 1998 Genetic Improvement of Cattle and Sheep. Farming Press: Ipswich, U.K.
777
778
Statham PTE, Green LE and Mendl M 2006 Development of a propensity to chew test
as a predictor of tail-biting in pigs. Bristol, U.K.,
779
780
Su G, Lund MS and Sorensen D 2007 Selection for litter size at day five to improve
litter size at weaning and piglet survival rate. Journal of Animal Science 85: 1385-1392
781
782
Svendsen PM, Hansen BK, Malmkvist J, Hansen SW, Palme R and Jeppesen LL
2007 Selection against stereotypic behaviour may have contradictory consequences for
25
783
784
the welfare of farm mink (Mustela vison). Applied Animal Behaviour Science 107: 110119
785
786
787
T-AP 1995 Standing committee of the european convention for the protection of animals
kept for farming purposes (T-AP): Recommendations concerning Domestic Fowl (Gallus
gallus).
788
789
T-AP 1999 Standing committee of the european convention for the protection of animals
kept for farming purposes (T-AP): Recommendations concerning fur animals.
790
791
T-AP 2005a Standing committee of the european convention for the protection of
animals kept for farming purposes (T-AP). Recommendations concerning pigs.
792
793
T-AP 2005b Standing committee of the european convention for the protection of
animals kept for farming purposes (T-AP): Recommendations concerning farmed fish.
794
795
796
Turner SP, D'Eath RB, Roehe R and Lawrence AB 2009a Selection against pig
aggressiveness at re-grouping; practical application and implications for long-term
behavioural patterns. Animal Welfare this issue:
797
798
799
Turner SP, Farnworth MJ, White IMS, Brotherstone S, Mendl M, Knap P, Penny P
and Lawrence AB 2006a The accumulation of skin lesions and their use as a predictor of
individual aggressiveness in pigs. Applied Animal Behaviour Science 96: 245-259
800
801
802
Turner SP and Lawrence AB 2007 Relationship between maternal defensive
aggression, fear of handling and other maternal care traits in beef cows. Livestock
Science 106: 182-188
803
804
805
806
Turner SP, Roehe R, D'Eath RB, Ison SH, Farish M, Jack MC, Lundeheim N,
Rydhmer L and Lawrence AB 2009b Genetic validation of skin injuries in pigs as an
indicator of post-mixing aggressiveness and the relationship with aggression under stable
social conditions. Journal of Animal Science in press:
807
808
809
810
Turner SP, Roehe R, Mekkawy W, Farnworth MJ, Knap PW and Lawrence AB
2008 Bayesian analysis of genetic associations of skin lesions and behavioural traits to
identify genetic components of individual aggressiveness in pigs. Behavior Genetics 38:
67-75
811
812
813
814
Turner SP, White IMS, Brotherstone S, Farnworth MJ, Knap PW, Penny P, Mendl
M and Lawrence AB 2006b Heritability of post-mixing aggressiveness in grower-stage
pigs and the relationship between aggressiveness and production traits. Animal Science
82: 615-620
815
816
817
Van Oers K, de Jong G, Van Noordwijk AJ, Kempenaers B and Drent PJ 2005
Contribution of genetics to the study of animal personalities: a review of case studies.
Behaviour 142: 1185-1206
26
818
819
820
Vinke CM, Eenkhoorn NC, Netto WJ, Fermont PCJ and Spruijt BM 2002
Stereotypic behaviour and tail biting in farmed mink (Mustela vison) in a new housing
system. Animal Welfare 11: 231-245
821
822
823
824
825
Wolf BT, McBride SD, Lewis RM, Davies MH and Haresign W 2008 Estimates of the
genetic parameters and repeatability of behavioural traits of sheep in an arena test.
Applied Animal Behaviour Science 112: 68-80
27
Behaviour
Social
Aggression
Poultry
Pigs
Sheep
Cattle
Selection line studies (Craig
0.17-0.46 (Løvendahl et
0.28-0.36 (Silva et al
et al 1965)
al 2005; Turner et al
2006)
Fur animals
2006b; Turner et al
2008; Turner et al
2009b)
Sociality
Selection line studies (Mills
0.02 – 0.39*
& Faure 1991)
(Wolf et al
2008)
Abnormal
Damaging
Feather pecking 0.11-0.38
Tail biting 0.05 (Breuer
Fur chewing 0.30
conspecifics
(Kjaer & Sorensen 1997;
et al 2005)
(Nielsen &
Rodenburg et al 2003);
Therkildsen 1995;
Selection line studies (Craig
cited by Malmkvist &
& Muir 1993; Buitenhuis &
Hansen 2001)
Kjaer 2008)
Stereotypy
Selection line studies (Mills
Selection line studies
et al 1985b)
(Hansen 1993a;
Jeppesen et al 2004)
Fear
of humans/
0.08 – 0.34 (Craig & Muir
0.38 (Hemsworth et al
0.02 – 0.39*
0.06-0.44
0.38 (Hansen 1993b;
handling ease
1989); Tonic immobility
1990)
(Wolf et al
(Beef, Le Neindre et
cited by, Malmkvist
selection line studies (Faure
0.03 – 0.17 (D'Eath et
2008)
al 1995; Phocas et al
& Hansen 2001)
& Mills 1998)
al 2009)
2006; Kadel et al
2006)
28
0.07 (Dairy, Pryce et
al 2000)
of novel objects or
tonic immobility selection
0.16 (Beilharz & Cox
places
lines (Mills & Faure 1991);
1967)
Open field 0.10 – 0.49
(Rodenburg et al 2003)
Reproductive
Maternal behaviour
0.01-0.08 (Grandinson
0.13 (Lambe et
0.06 – 0.09
et al 2003; Løvendahl et
al 2001)
(defensive
al 2005)
aggression, reviewed
by Burrow 1997)
Table 1: Examples of evidence for a genetic component of behaviour traits in farm animals. Evidence of successful selection
experiments or estimates of heritability (h2) from pedigree studies are given. Where a range of values is reported, this reflects
both the use of multiple variables or test ages within one study, and differences across studies. *This study is difficult to
classify as it recorded behaviour in a test of conflicting motivations (avoid a human vs. seek flock mates).
29